Why labor shortages, digital transformation, and productivity are parts of the same operating problem
Manufacturers often treat labor, digital transformation, and productivity as separate programs. Michael Carroll argues that they are connected by the same operating problem: how knowledge moves, how decisions are made, and whether people have the authority and experience to act.
In this episode of MetaPod, Ron Crabtree and Carroll examine what happens when companies outsource learning, digitize old decision paths, and ask a smaller workforce to manage more complexity. The discussion points toward an enterprise that can sense a problem and respond with less delay.
Outsourcing can break the learning loop
Entry-level work does more than produce an immediate output. It teaches employees what normal looks like and, just as important, what wrong looks like. People build judgment through repeated exposure, then carry that judgment into roles with greater responsibility.
When companies outsource those early roles without another way to build experience, they interrupt the learning loop. The near-term labor saving can be offset by more meetings, slower decisions, and weaker problem solving. Over time, the organization may lose the people who can recognize trouble before it becomes downtime.
AI can create the same mistake if it removes work without preserving how people learn. Keeping a human in the loop is not enough when that person no longer has the experience to challenge the system's answer.
A labor shortage can be a knowledge architecture problem
Carroll describes plants where experienced generalists kept work flowing across unit operations. Specialists with decades of experience could recognize unusual conditions before a failure spread. As tenure fell, unplanned downtime increased because newer employees had not seen enough exceptions to predict what would happen next.
The problem was not intelligence or effort. The operating model required judgment that took years to form. If a facility depends on people who need 25 years of experience, leaders must ask whether those people will exist when the asset reaches its next stage of life.
Companies should capture expertise from exceptions, not only from standard procedures. Procedures describe the expected path. The harder knowledge lives in the moments when the process changes state, conditions conflict, and an experienced employee chooses what to do.
Digitizing a slow decision keeps it slow
Technology can move information faster while leaving the decision process untouched. If an engineering approval still waits in a queue, automating an earlier step may shift the constraint without improving the overall result.
Carroll calls the delay between a signal and an authorized action decision latency. It appears in approval queues, alignment meetings, unclear ownership, and requests for permission that could have been defined in advance. That delay costs margin because the business keeps waiting while the problem continues.
Measure the full distance from signal to changed outcome. A faster report matters only if it helps someone act sooner and improves the result.
Start with the decision
Before selecting technology, define the decision the business needs to make, the outcome it should produce, and the person with the best knowledge and perspective to own it. Then identify where the work requires inference and where it requires permission.
Routine work can be standardized. Non-routine work needs room for judgment. AI can handle much of the information transfer around a decision, but a change to a record, asset, or operating state carries responsibility. That responsibility needs a clear chain of custody.
Design permission before the signal arrives
Many organizations wait for a problem and then gather people to decide who can act. A better approach defines authority in advance. Guardrails can specify which conditions allow an employee or AI system to proceed, when a second approval is required, and which cases must stop.
A decision ledger can record the signal, action, result, and reason. The ledger helps the organization see whether its rules improve speed and outcomes. It also exposes repeated exceptions that deserve a new standard or a different allocation of authority.
Collaboration should have a purpose. People who own the decision need to participate. People who only need awareness can receive the information without joining another meeting.
Put people where responsibility changes state
Carroll separates information transfer from changes that carry responsibility. AI can summarize, route, compare, and prepare information. People remain important where the decision changes the state of a product, process, account, or asset and someone must be accountable for the result.
This distinction helps leaders use scarce expertise well. Instead of spreading experienced employees across routine coordination, the company places them at the points where judgment and permission matter most. AI supports those employees with context and handles the lower-risk transfer work around them.
Connect and learn more
The one degree enterprise is less about a new tool than a shorter path between evidence and action. It preserves learning, assigns ownership, and gives people the context and authority to respond before a problem grows.
To learn more, connect with Michael Carroll on LinkedIn. You can also connect with Ron Crabtree on LinkedIn or contact the MetaPod team at MetaExperts.com.